Background: Alcohol dependence (AD) is a biologically, genetically based disease, yet the majority of clinically accepted treatments are behaviorally or psychosocially based. Despite the initial success of these treatments, 40-70% of patients relapse within the first 12 months after initiating treatment. Neuropharmacology of alcohol and prazosin: Emerging pre-clinical evidence shows that noradrenergic systems are involved in brain processes relevant to AD, such as arousal, reinforcement, and stress responsivity. However, virtually no work to date has attempted to translate this knowledge into clinically effective biological interventions. We have adopted the novel, promising strategy of reducing adrenergic activity by blocking noradrenaline binding to post-synaptic a1 receptors via the non-selective, a1 antagonist, prazosin. Preclinical studies have demonstrated that prazosin decreases reinstatement of alcohol consumption, and preliminary clinical pilot data suggest that prazosin reduces alcohol use in humans with AD. Prazosin, FDA approved to treat hypertension, typically has few side effects, and is inexpensive. Design: Randomized double-blind placebo-controlled clinical trial. Participants: 120 AD individuals (25%women) with stated goal to abstain from alcohol use and without PTSD. Intervention: Either prazosin titrated per study protocol or matched placebo for 12 weeks with Medical Management (MM) based on the COMBINE Study procedures and a final study visit one month after medication discontinuation. Measures: The primary outcomes are alcohol use during the 12-week medication phase of the study and reports of craving during the same time period. Daily, prompted Interactive Voice Response (IVR) telephone monitoring will be done throughout the 12-week medication phase of the study to assess the primary outcomes and to provide information on affect and medication adherence. Such daily monitoring provides more accurate reports of alcohol use than standard retrospective outcome measures. Furthermore, since alcohol craving and use can occur precipitously with antecedent events promptly forgotten, daily monitoring will enhance the capacity to evaluate the relationship between these phenomena. Analyses: Hierarchical linear modeling to test for main effects of prazosin+MM vs placebo+MM on alcohol craving and use over time, and to evaluate whether reductions in craving mediate the effect of prazosin.